Re: Mixed-model admission tests
Dennis Cottel (dennis@nosc.mil)
Mon, 16 Dec 1996 14:40:20 -0800
Without responding in detail to Kelvin Nilsen's notes on scheduling, I'd
like to suggest that much of this detail may not really belong in MPI --
which is for sending messages, after all, not scheduling processes. Let the
application (in conjunction with the platform's scheduling API) do the
detailed task scheduling while MPI/RT just provides certain bandwidth
guarantees for the messages.
Dennis Cottel, dennis@nosc.mil, (619) 553-1645
NCCOSC, RDT&E Division (NRaD), San Diego, CA